[Qirg] E91 (was Re: Measurements in link layer service)

[Rodney Van Meter <[e-mail]>]

> 
> 2. On might argue that in this case it is not really BB84 but more like E91, since BB84 concerns preparing a state and sending it to another node. However, note that preparing an entangled state and measuring on one side is almost equivalent to sending a random state, except for how the state that is received is sampled.

E91 (https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.67.661 <https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.67.661>, findable in a number of places online) is an *idea* rather than a protocol: that key generation could be done over Bell pairs in which both ends measure in a randomly-chosen basis (selected from a small set) and then post-processed.  Some are used for verifying CHSH violation (proving that you have two-party entanglement, which puts a strict limit on how much information can be leaking to an eavesdropper) and some are used to generate the key material itself.

But E91 is nothing at all like a complete protocol! BB84, of course, isn’t a complete protocol in and of itself, but it does go a step further in laying out the sequence of operations, and lots of work has been done to build a complete protocol suite around it (much of which isn’t standardized yet).  But E91 is waaay farther behind in terms of moving from concept to protocol.

—Rod